• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 31 subscribers
  • Views 604 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

IWRL6432AOP: Issues with Tracking Data not keeping up with Point Cloud data

Jack
Mastermind 9145 points
Part Number: IWRL6432AOP
Other Parts Discussed in Thread: IWRL6432

Tool/software:

Hi L6432AOP Champ !

Here's the test environment to start with.

[RADAR SENSOR] IWRL6432AOP EVM

[SDK] MMWAVE_L_SDK_05_05_00_02

[Radar Toolbox] 2.20.00.05

For the L6432AOP test, They further compared the cfg files from the IWRL6432 EVM and the IWRL6432AOP EVM and found the main differences as shown below.

IWRL6432 EVM  chirpCommCfg 16 0 0 128 4 28 0
IWRL6432AOP EVM  chirpCommCfg 15 0 0 128 1 28 2

After checking, it seems to be a difference in the BPM/TDM mode, so I changed the mode from TDM(1) to BPM(4) and tested it, and I saw a big improvement over the previous contrast, although some of the tracking data does not follow the point cloud.

However, after running the program and loading the cfg, I noticed that there is a WARNING in the log like below.
Warning! BPM is not supported on this device

It's hard to tell if the improvement is due to the change to BPM mode. Or is there something else going on?

Please check this out.

Thanks.

Best Regards, 

Jack

  • 0
    TI__Mastermind 23177 points

    Hello Jack,

    The warning only shows on the AOP, as due to its very compact antenna pattern the AOP requires some differences in configuration. You can run BPM, but it will likely be saturated with a high amount of false positives. If you are looking for better detection, I recommend for AOP to keeping it in TDM and then adjusting CFAR thresholding parameters.

    Best Regards,

    Pedrhom

  • 0 in reply to Pedrhom Nafisi
    Prodigy 80 points

    High Pedrhom,

    The current issue is that the target data cannot keep up with the point cloud data.

    As the distance between the point cloud and the target increases, the target data eventually gets deallocated and reallocated as new target data.


    The video contains the test results and includes the following content.

     - Initially, the target data is well allocated to the point cloud.

     - As the object moves, the point cloud also shifts, and the target data follows accordingly.

     - At a certain point, the target data can no longer keep up with the point cloud, and eventually, the target is deallocated.

     - New target data is then assigned to the point cloud.

    This problem is severe in TDM mode, while it's somewhat less prominent in BPM mode.

    Please check and advise.

    Thank you.

  • 0 in reply to SeungCheol Kim
    TI__Mastermind 23177 points

    Hello,

    This is a bit of a different issue, in the sense that the solution is to change the AOP's tracker parameters.

    In the configuration file, I would increase the maxAcceleration parameter values, the maxRadialVelocity field in the trackingCfg parameter, and the limit velocity field within gatingParam.

    For a tracker tuning guide, please refer to the following

    https://dev.ti.com/tirex/explore/node?a=1AslXXD__2.20.00.05&node=A__AI4900ZWHa65Tb4Eso4FzQ__radar_toolbox__1AslXXD__2.20.00.05

    Best Regards,

    Pedrhom

  • 0 in reply to Pedrhom Nafisi
    Prodigy 80 points

    Hello,

    I made modifications as per the tracker tuning guide and conducted a test with the following adjustments:

    In the [cfg] TrackingClassification_MidBw.cfg({MMWAVE_SDK5_INSTALL}\examples\mmw_demo\motion_and_presence_detection\profiles\xwrL64xx-aop) file, I only modified the following:

    maxAcceleration 1.0 1.0 0.2
    gatingParam 3 2.5 2.5 2.5 8

    Compared to before, it can be seen that the target data is now better able to follow the point cloud without losing track.

    However, as shown in the video below, when a person moves at a constant speed, the target data does not follow smoothly. Instead, it fails to keep up properly and then suddenly moves rapidly to catch up.

    This issue did not occur with the default configuration file on the IWRL6432EVM, but it seems to occur only with the IWRL6432AOP.

    Is there any difference between the two EVMs other than the antenna configuration?

    Please check and confirm.

    Sincerely,

    SeungCheol Kim

  • 0 in reply to SeungCheol Kim
    TI__Mastermind 23177 points

    Hello SeungCheol Kim,

    The AOP has slightly worse point cloud detection/accuracy performance due to its very compact design. The antenna configuration may be causing different velocity readings. Try increasing the maxRadialVelocity field in the trackingCfg parameter as well, but I do agree the track should be keeping up with the points. I do not believe this is particularly the sensor's fault as we can clearly see enough points that the tracker should be kept on target, it just needs some tuning to adhere to the differences.

    Best Regards,

    Pedrhom